1. Field of the Invention
The present invention relates to a wireless communication system which improves reliability and throughput of network, and a retransmission timeout determining method used for the same. More particularly, the present invention relates to a wireless communication system, a server, a data relay apparatus, a base station apparatus, a mobile terminal which improves reliability and throughput of network, a retransmission timeout (RTO) determining (establishing) method used for the same and its program.
2. Description of the Related Art
Conventionally, in a wireless communication system, there is a system in which a wireless communication section and a wire communication section co-exist and a data is transmitted by using TCP as a communication protocol. For example, a system, in which a portable terminal and a server carry out a two-way communication through a base station and a data relay apparatus, employs a following procedure. A receiving end transmits an acknowledgement (ACK) with regard to a received data to a transmitting end, and if the acknowledgement is not returned within a certain time after the transmitting side transmits the data, the transmitted data is retransmitted.
The retransmission timeout is the time to carry out the retransmission of the transmitted data, if the acknowledgement with regard to the transmitted data can not be received. The TCP employs a method of determining the optimal time for the retransmission timeout based on a round trip time (RTT) that is a temporal difference between a transmission time of the data and a reception time of the acknowledgement. With regard to the establishing method of this retransmission timeout, a method defined in RFC2988 widely known as a Jacobson method is typical.
However, this method uses an equation introduced by a queue theory assuming a wire network. It does not assume a network in which an error rate and a communication rate are largely varied, such as a wireless network. Thus, there is a problem that the retransmission timeout is not proper for the wireless network.
As a method of solving this problem, a following method is proposed in the Japanese Laid Open Patent Application (JP-A 2002-330168). Firstly, a queueing time data at a link layer level of the base station in the wireless communication section between the base station and the portable terminal is reported from the portable terminal to the server, as a TCP option in an SYN packet at a time of a TCP connection establishment. Next, the server analyzes the TCP option in the SYN packet and obtains the queueing time data at the link layer level of the base station. Then, the server adds this queueing time data to the retransmission timeout at a time of a TCP packet transmission. The above-mentioned queueing time data is the time equal to n (n is a positive integer) multiples of a retransmission interval T from the base station to the portable terminal at the link layer level.
Also, another establishing method of a retransmission timeout is disclosed in Japanese Laid Open Patent Application (JP-A-Heisei 8-8995). Firstly, by recording a transmission time of a transmission frame and monitoring a reception check frame, a response time is determined. Next, a dynamic retransmission timer is dynamically updated on the basis of the maximum and the minimum value of a response time, a retransmission flag and an optimal value in the past communications. Then, by using the optimal reception check frame queue timeout value, a retransmitting process is executed.
In conjunction with the above description, Japanese Laid Open Patent Application (JP-A 2002-135277) discloses a method of controlling a time-out of a communications protocol. The method of controlling a time-out of a communications protocol, includes a step which delays a part of a plurality of messages transmitted on a communication channel. The messages may be the message which should be transmitted or received.
In conjunction with the above description, Japanese Laid Open Patent Application (JP-A-Heisei 10-126459) discloses a method for setting a time-out of the transmitting and responding in the LAN equipment. The method for setting a time-out of the transmitting and responding in the LAN equipment which transmits data with a TCP/IP protocol, includes a echo demand unit, a timer, a operation unit, a time storage unit and a time-out control unit. The echo demand unit performs an echo demand of a data of a specified quantity. The timer measures a delay time form a start of data transmission to a start of data reception. The operation unit computes a retransmission time-out based on the delay time. The time storage unit stores the time-out. The time-out control unit controls the above units.
In conjunction with the above description, Japanese Laid Open Patent Application (JP-A 2002-281106) discloses a data communication method for performing data communication. The data communication method for performing data communication using a transport control protocol (TCP) among a plurality of communication terminals through a communication network containing wireless communication lines, is characterized in that a retransmission time-out in case of carrying out data transmission from the communication terminal is longer than a recommended value of a retransmission time-out time in case of constituting the communication network only by wire communication lines.
In conjunction with the above description, Japanese Laid Open Patent Application (JP-A 2001-160842) discloses a data communication system. The data communication system which transmits and receives data with an application server using the same protocol through one of transmission channels of a wire network, a wireless network and both of them, includes a first table and a retransmission timer setting unit. The first table stores a relation among a service identifier to an application server, a transmission address, a kind of a communication networks, and a processing time of the application server. The retransmission timer setting unit which calculates a processing time of the application server based on a service identifier to an application server, a transmission address and a kind of a communication networks during the data transmission to the application server with reference to the first table, and set up the processing time as a retransmission time.
In conjunction with the above description, Japanese Laid Open Patent Application (JP-A-Heisei 06-252978) discloses a network parameter automatic adjusting apparatus. The network parameter automatic adjusting apparatus includes a transmission parameter table, a transmission control means, a parameter adjustment control means and a parameter adjustment means. The transmission parameter table is installed in each of a plurality of computers connected on a network and describes a data retransmission interval and a data retransmission count for a reply queue time and a time-out of the reply queue time, as parameters. The transmission control means performs a data transmission with another computer through the above-mentioned network based on the parameters described in the transmission parameter table. The parameter adjustment control means makes the transmission control means transmit a frame for a transmission performance check to a transmission control means of the other computer, and measures a response period from transmitting the frame to receiving a response frame to the performance check frame transmitted by the other computer, and makes the transmission control means transmit another response frame to the performance check frame transmitted by the other computer. The parameter adjustment means corrects each parameter described in the transmission parameter table based on a response time of the performance check frame measured by the parameter adjustment control means.
In case of the technique of the JP-A 2002-330168, the above-mentioned conventional retransmission timeout establishing method establishes the retransmission timeout in accordance with the queueing time data at the link layer level of the base station. Thus, in the system in which the base stations of different manufacturers co-exist, there may be a possibility that the magnification n and the retransmission interval T at the link layer level from the base station to the portable terminal is different for each base station. Hence, this causes the problem that it is impossible for the conventional method to treat this case. Also, the technique of the JP-A 2002-330168 has a problem that since it uses the TCP option of the SYN packet, it can not be applied to a system or a network if the TCP option can not be used.
The technique of the JP-A-Heisei 8-008995 has also a problem that if the maximum and the minimum response times at the time of the past communications are updated, the interval between the maximum response time and the minimum response time becomes large. If the error rate and the communication rate are largely varied in the network, such as the wireless network. The retransmission timeout is not proper for the communication.